Browsing by Author "Liu, Gi-Ren"

Now showing 1 - 4 of 4

Diffuse to fuse EEG spectra - Intrinsic geometry of sleep dynamics for classification
(2020-01-01)
Liu, Gi-Ren
;
Lo, Yu-Lun
;
Malik, John
;
Sheu, Yuan-Chung
;
Wu, Hau-Tieng
We propose a novel algorithm for sleep dynamics visualization and automatic annotation by applying diffusion geometry based sensor fusion algorithm to fuse spectral information from two electroencephalograms (EEG). The diffusion geometry approach helps organize the nonlinear dynamical structure hidden in the EEG signal. The visualization is achieved by the nonlinear dimension reduction capability of the chosen diffusion geometry algorithms. For the automatic annotation purpose, the support vector machine is trained to predict the sleep stage. The prediction performance is validated on a publicly available benchmark database, Physionet Sleep-EDF [extended] SC* (SC = Sleep Cassette) and ST* (ST = Sleep Telemetry), with the leave-one-subject-out cross validation. When we have a single EEG channel (Fpz-Cz), the overall accuracy, macro F1 and Cohen's kappa achieve 82.72%, 75.91% and 76.1% respectively in Sleep-EDF SC* and 78.63%, 73.58% and 69.48% in Sleep-EDF ST*. This performance is compatible with the state-of-the-art results. When we have two EEG channels (Fpz-Cz and Pz-Oz), the overall accuracy, macro F1 and Cohen's kappa achieve 84.44%, 78.25% and 78.36% respectively in Sleep-EDF SC* and 79.05%, 74.73% and 70.31% in Sleep-EDF ST*. The results suggest the potential of the proposed algorithm in practical applications. (C) 2019 Elsevier Ltd. All rights reserved.
Modeling Mobile Ticket Dispenser System With Impatient Clerk
(2016-12)
Liu, Gi-Ren
;
Lin, Phone
;
Lin, Yi-Bing
;
資訊工程學系
;
Department of Computer Science
The mobile ticket dispenser system (MTDS) allows customers to remotely draw tickets for service orders anywhere through a mobile handset. In our previous work, the MTDS was applied to a restaurant scenario in which both clerks and customers are patient, i.e., once a mobile ticketing (MT) customer remotely draws the ticket, his request can be served by the clerk when the clerk is available, regardless of when the customer arrives at the restaurant. In this paper, the MTDS is applied to a post office scenario in which the customers are patient, but the clerk is impatient since the original ticket drawn by theMT customer may be invalid if he/she does not arrive at the post office before his/her turn. In this case, the behavior of the MT customer is the same as the so-called in-person ticketing customer who needs to draw a ticket in person when he/she arrives at the service counter. We propose an analytical model to derive the probability that an MT customer misses his/her turn when he/she arrives at the post office. A discrete-event simulation model is developed to investigate the performance of the predicted time adjustment mechanism for the MTDS. We also use real data collected at a post office to observe the queuing behavior. Our study provides guidelines for arranging the time for an MT customer to arrive at the MTDS server.
Optimal Threshold Policy for In-Home Smart Grid with Renewable Generation Integration
(2015-04-01)
Liu, Gi-Ren
;
Lin, Phone
;
Fang, Yuguang
;
Lin, Yi-Bing
;
資訊工程學系
In-home Smart Grid (SG), the integration of Renewable Power Systems (RPSs) with Conventional Power Systems (CPSs), calls for cost-effective management for the electricity usages of end users\' household appliances. In this paper, by taking the charging process of RPSs and multiple types of household appliances in to consideration, we have developed analytical models to characterize the electricity cost in the in-home smart grid. Based on these models, we formulate the electricity cost minimization problem as a finite-horizon continuous-time Markov decision process (CTMDP), from which we obtain a threshold policy to minimize the cost. Numerical results show that the threshold policy can manage the electricity usage very effectively.
Save Muscle Information-Unfiltered EEG Signal Helps Distinguish Sleep Stages
(2020-04-01)
Liu, Gi-Ren
;
Lustenberger, Caroline
;
Lo, Yu-Lun
;
Liu, Wen-Te
;
Sheu, Yuan-Chung
Based on the well-established biopotential theory, we hypothesize that the high frequency spectral information, like that higher than 100Hz, of the EEG signal recorded in the off-the-shelf EEG sensor contains muscle tone information. We show that an existing automatic sleep stage annotation algorithm can be improved by taking this information into account. This result suggests that if possible, we should sample the EEG signal with a high sampling rate, and preserve as much spectral information as possible.